On Tue, Nov 2, 2021 at 8:14 AM Suren Baghdasaryan [off-list ref] wrote:

On Tue, Nov 2, 2021 at 12:58 AM Michal Hocko [off-list ref] wrote:

quoted

On Mon 01-11-21 08:44:58, Suren Baghdasaryan wrote:
[...]

quoted

I'm with you on this one, that's why I wanted to measure the price we
would pay. Below are the test results:

Test: https://lore.kernel.org/all/20170725142626.GJ26723@dhcp22.suse.cz/ (local)
Compiled: gcc -O2 -static test.c -o test
Test machine: 128 core / 256 thread 2x AMD EPYC 7B12 64-Core Processor
(family 17h)

baseline (Linus master, f31531e55495ca3746fb895ffdf73586be8259fa)
p50 (median)   87412
p95                  168210
p99                  190058
average           97843.8
stdev               29.85%

unconditional mmap_write_lock in exit_mmap (last column is the change
from the baseline)
p50 (median)   88312     +1.03%
p95                  170797   +1.54%
p99                  191813   +0.92%
average           97659.5  -0.19%
stdev               32.41%

unconditional mmap_write_lock in exit_mmap + Matthew's patch (last
column is the change from the baseline)
p50 (median)   88807      +1.60%
p95                  167783     -0.25%
p99                  187853     -1.16%
average           97491.4    -0.36%
stdev               30.61%

stdev is quite high in all cases, so the test is very noisy.
The impact seems quite low IMHO. WDYT?

Results being very noisy is what I recall as well. Thanks!

I believe, despite the noise, the percentiles show that overall we do
not noticeably regress the exit path by taking mmap_lock
unconditionally.
If there are no objections, I would like to post a patchset which
implements unconditional locking in exit_mmap() and process_madvise()
calling __oom_reap_task_mm() under protection of read mmap_lock.
Thanks!

Discussing how the patch I want to post works for maple trees that
Matthew is working on, I've got a question:

IIUC, according to Michal's post here:
https://lore.kernel.org/all/20170725154514.GN26723@dhcp22.suse.cz (local),
unmap_vmas() can race with other mmap_lock read holders (including
oom_reap_task_mm()) with no issues.
Maple tree patchset requires rcu read lock or the mmap semaphore be
held (read or write side) when walking the tree, including inside
unmap_vmas(). When asked, he told me that he is not sure why it's
currently "safe" to walk the vma->vm_next list in unmap_vmas() while
another thread is reaping the mm.
Michal (or maybe someone else), could you please clarify why
unmap_vmas() can safely race with oom_reap_task_mm()? Or maybe my
understanding was wrong?
Thanks,
Suren.

quoted

--
Michal Hocko
SUSE Labs